• 한국컴퓨터정보학회논문지
• /
• 제24권1호
• /
• pp.137-143
• /
• 2019

As the performance of computers increases, the use of deep learning, which has faced technical limitations in the past, is becoming more diverse. In many fields, deep learning has contributed to the creation of added value and used on the bases of more data as the application become more divers. The process for obtaining a better performance model will require a longer time than before, and therefore it will be necessary to find an optimal model that shows the best performance more quickly. In the artificial neural network modeling a tuning process that changes various elements of the neural network model is used to improve the model performance. Except Gride Search and Manual Search, which are widely used as tuning methods, most methodologies have been developed focusing on heuristic algorithms. The heuristic algorithm can get the results in a short time, but the results are likely to be the local optimal solution. Obtaining a global optimal solution eliminates the possibility of a local optimal solution. Although the Brute Force Method is commonly used to find the global optimal solution, it is not applicable because of an infinite number of hyper parameter combinations. In this paper, we use a statistical technique to reduce the number of possible cases, so that we can find the global optimal solution.

• 동력기계공학회지
• /
• 제3권4호
• /
• pp.69-78
• /
• 1999

In the single-input and single-output system, the parameter of plant is scalar polynomial, but in the multiple input and multiple output, it accompanies, being matrix polynomial, the consideration of observable controlability index or problems non-commutation in matrix polynomial as well as degree, and it is more complex to deal with. Therefore, it is thought that a full research on the single-input and single-output system is not sufficient. This paper proposes that problems of minimum variance self-tuning regulator by using numerical calculation example of multivariable system and pole assignment self-tuning regulator.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2008년도 춘계학술대회 논문집
• /
• pp.469-470
• /
• 2008

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1994년도 Proceedings of the Korea Automatic Control Conference, 9th (KACC) ; Taejeon, Korea; 17-20 Oct. 1994
• /
• pp.449-454
• /
• 1994

This paper describes new PID control methods based on the fuzzy logic. PID gains are retuned after evaluating control performances of transient responses in terms of performance features. The retuning procedure is based on fuzzy rules and reasoning accumulated from the knowledge of experts on PID gain scheduling. For the case that the retuned PID gains result in worse CLDR (characteristics of load disturbance rejection) than the initial gains, an on-line tuning scheme of the set-point weighting parameter is, proposed. This is based on the fact that the set-point weighting method efficiently reduce either overshoot or undershoot without any degradation of CLDR. The set-point weighting parameter is adjusted at each sampling instant by the fuzzy rules and reasoning. As a result, better control performances were achived in comparison with die controllers tuned by the Z-N (Ziegler-Nichols) parameter tuning formula or by the fixed set-point weighting parameter.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2000년도 추계학술대회 학술발표 논문집
• /
• pp.195-198
• /
• 2000

Sliding mode control method is popularly used for robustness to distrurbance and variance of systems internal parameter. However, one of the serious problem of this method is Chattering which occurs in neighborhood of sliding manifold. Another problem is that we cannot expect robustness before system starts sliding mode. A new tuning method of sliding manifold which changes the parameter of sliding manifold dynamically using Wavelet Neural Network is proposed in this paper. We can expect the better performance in sliding mode control by the wavelet neural networks excellent property of approximating arbitrary function for multi-resolution analysis and decrease chattering drastically.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1987년도 한국자동제어학술회의논문집; 한국과학기술대학, 충남; 16-17 Oct. 1987
• /
• pp.32-35
• /
• 1987

An adaptive control scheme has been recognized as an effective approach for a robot manipulator to track a desired trajectory in spite of the presence of nonlinearties and parameter uncertainties in robot dynamic models. In this paper, an adaptive control scheme for a robot manipulator is proposed to design the self-tuning controller which combines the pole placement with the extended linearized perturbation model. And this control scheme has two components: a feadforward control and a feedback compensation control. Based on this, the controller is demonstrated by the simulation about position control of a three-link manipulator with payload and parameter uncertainty.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 2013년도 전력전자학술대회 논문집
• /
• pp.315-316
• /
• 2013

This paper presents a parameter tuning method of a LLC resonant converter for a wireless charging circuit. A switched-capacitor is used to change the resonant frequency of the resonant circuit. The simulation results verify that the efficiency of the power transfer can be improved by a duty control of the switched-capacitor for various values of the coupling coefficient.

• 대한전기학회논문지
• /
• 제37권9호
• /
• pp.655-662
• /
• 1988

An adaptive control scheme has been recognized as an effective approach for a robot manipulator to track a deired trajectory in spite of the presence of nonlinearies and parameter uncertainties in robot dynamic models. In this paper, an adaptive control scheme for a robot manipulator is proposed to design the self-tuning controller which controls the extended linearized perturbaton model via the pole placement, and this control. The feasibility of the controller is demonstrated by the simulation about position control of a three-link manipulator with payload and parameter uncertainty.

• 대한전기학회논문지
• /
• 제40권2호
• /
• pp.183-191
• /
• 1991

This paper develops a pole-zero assignment self-tuning regulator utilizing the method of a neural network in the plant parameter estimation. An approach to parameter estimation of the plant with a Hopfield neural network model is proposed, and the control characteristics of the plant are evaluated by means of a simulation for a second-order linear time invariant plant. The results obtained with those of Exponentially Weighted Recursive Least Squares(EWRLS) method are also shown.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2005년도 추계학술대회 학술발표 논문집 제15권 제2호
• /
• pp.423-426
• /
• 2005

In this paper, we proposed a PID tuning algorithm by the fuzzy set theory to improve the performance of the PID controller. The new tuning algorithm for the PID controller has the initial value of parameter Kp, $\tau_{I}$, $\tau_{D}$. by the Ziegler-Nichols formula that uses the ultimate gain and ultimate period from a relay tuning experiment. We will get the error and the error rate of plant output corresponding to the initial value of parameter and fnd the new proportion gain(Kp) and the integral time ($\tau_{I}$) from fuzzy tuner by the error and error rate of plant oueut as a membership function of fuzzy theory. This fuzzy auto tuning algorithm for PID controller considerably reduced the overshoot and rise time as compared to any other PID controller tuning algorithms. And in real parametric uncertainty systems, it constitutes an appreciable improvement of performance. The significant property of this algorithm is shown by simulation